Abhigya Goel

Hardware · Embedded · Robotics

Hey, I'm Abhigya.

I'm a junior (unc) at USC, working on software that meets the physical world.

I'm happiest when code leaves the screen; I love hardware, drones, robotics, and embedded systems — if it moves in the real world, I've probably worked on it. That sentence wasn't AI.

I also care a lot about how websites look, music, I cook, I keep up with what's happening in the world (yikes), and I hang out with my friends.

This summer I worked embedded and firmware at RTX. Before that I worked robotics for NASA and a whole bunch of other cool side quests. Mostly, I just like building. Right now I'm making a personal JARVIS on a Jetson, and every day my own hardware internship scraper texts me (a lot).

I know I'm supposed to “own a niche” but really I just like making anything that does something real or helps out. Anyways, thanks for stopping by and scroll around a little.

Present

  • Developing embedded firmware test platform in Python & C on Linux, simulating ESP32 firmware across dozens of emulations, scaling multi-node distributed system validation to 20+ node topologies on a single host, reducing need for physical hardware.
  • Simulating ESP-NOW and Wi-Fi network behavior across configurable topologies, generating Wireshark-readable packet traces.
  • Debug embedded systems & student labs in weekly office hours across Linux hosts and Raspberry Pi targets for 100+ students.

Previously On

  • Developed real-time embedded C++ & Python on Linux/RTOS for radar signal processing firmware, integrating and validating hardware-software interfaces across 9 subsystems for AN/TPY-2 missile defense radar under Secret clearance.
  • Automated embedded firmware regression and hardware integration testing across 8 subsystem repositories via Python and Jenkins, verifying cross-component interface integrity and build artifact delivery.
Trojan Debate SocietyDirector of Curriculum & Community Servicecurriculum + community service lead2025-2026
USC Advanced Robot CombatElectrical Engineering Leadled mini combat-robot electronics2025-2026
  • Built Python data pipeline automating REST API scraping of 50+ articles daily and BigQuery ingestion of federal AI use and policy news, enabling real-time tracking of U.S. government AI adoption across 20+ agencies.
  • Built SLM-powered NLP deduplication using sentence-transformers & cosine similarity to cluster 15,000+ datapoints by topic, reducing 30% redundancy, containerized with Docker & deployed on Google Cloud Run with automated daily ETL.
  • Developed multi-node ROS2 pipeline processing quadruped robot gait telemetry in real-time, estimating lunar terrain stiffness using forward kinematics and Jacobian-based force calculations, presented to NASA Ames researchers.
  • Built Foxglove visualization for live terrain stiffness heatmaps with uncertainty overlays from custom ROS2 message types.
  • Designed 3-month robotics curriculum teaching Arduino programming, circuit design, and sensor integration.
  • Taught 15+ students to build and code autonomous robots using motor control, ultrasonic sensing, and decision logic.

Projects

  • Built sub-$1000 autonomous drone on Raspberry Pi with Navio2 flight controller, designing custom mounts & power distribution across 5 subsystems interfacing LiDAR, camera, and IMU over UART, I2C, & SPI with PWM ESCs.
  • Developed ROS (C++) collision avoidance using LiDAR and potential field navigation, validating flight across 30+ field tests and diagnosing altitude oscillation and sensor drift faults from MAVLink telemetry logs.
  • Built & deployed CNN for human detection using TensorFlow, quantizing onto Google Edge TPU for 450% inference speedup.

ISEF

  • 6 parallel Claude agents handle intake, clinical reasoning, and live form analysis via Claude Vision over MediaPipe.
  • ElevenLabs voice intake and spoken coaching; per-rep tracking persisted in Supabase.

DemoGitHub

  • Prototyped 3D-printed assistive wearable in Fusion 360 with NVIDIA Jetson Orin Nano, utilizing OpenCV FFT algorithms and SciPy Digital Signal Processing to regulate incoming light/sound to reduce overstimulation for neurodivergent individuals.
  • Integrated GPT-3.5 with sensor telemetry to automate Arduino actuators via UART on real-time brightness/audio thresholds, exposed through a custom Flask REST API to a React Native companion app for remote, user-configurable interventions.

GitHub

  • Scrapes hardware-internship postings and texts me the moment new ones go up.

GitHub

  • Voice assistant on an NVIDIA Jetson that edits and redeploys its own source.

GitHub

  • Skill-based pairing and attendance for 60+ debaters biweekly, with live Firebase sync across devices.

Live

  • Built 4-node distributed embedded system (2x Arduino + Raspberry Pi) decoding real-time Morse via Random Forest classifier over USB serial, relaying decoded messages via MQTT with live Flask dashboard.
  • Wrote embedded Arduino firmware capturing microsecond-precision sensor input and driving LCD, buzzer, and LED device drivers over a serial command protocol, deploying pure-numpy ML inference on Raspberry Pi.

DemoGitHub

  • Modeled combat robot chassis in SolidWorks, managing mechanical tolerances to fit motors/electronics in 3D-printed frame.
  • Engineered power distribution harness, soldering LiPo, safety switch, and 5V BEC to 3 brushless ESCs and a 4-channel receiver.

Awards & Honors

2nd Overall: Anthropic SoCal Claude Hackathon2026
1st Place Hardware: a16z x USC AI for Good Buildathon2025
Regeneron ISEF Finalist: 5 of 400 from RSEF2024
IEEE Technology Innovation Award: 20242024
USC Presidential & Viterbi Scholarships: $230K+Present